Effect of Reducing Allochthonous P Load on Biomass and Alkaline Phosphatase Activity of Phytoplankton in an Urbanized Watershed, Michigan

Lake and Reservoir Management, 29:116–125, 2013 C Copyright by the North American Lake Management Society 2013 ISSN: 1040-2381 print / 2151-5530 online DOI: 10.1080/10402381.2013.800173

Effect of reducing allochthonous P load on biomass and alkaline phosphatase activity of phytoplankton in an urbanized watershed, Michigan

John T. Lehman,∗ Jonathan P. Doubek, and Elliot W. Jackson Department of Ecology and Evolutionary Biology, Natural Science Building, University of Michigan, Ann Arbor, MI 48109-1048

Abstract

Lehman JT, Doubek JP, Jackson EW. 2013. Effect of reducing allochthonous P load on biomass and alkaline phosphatase activity of phytoplankton in an urbanized watershed, Michigan. Lake Reserv Manage. 29:116–125.

Phosphorus (P) concentrations have declined in the Huron River in southeast Michigan subsequent to first a municipal ordinance, and later state legislation, restricting the application of commercial fertilizers containing phosphate. One management objective of the environmental policy was to reduce eutrophication of surface waters and nuisance blooms of cyanobacteria, particularly in large impoundments along the Huron River. Alternative management theory contended that nuisance algal conditions resulted from internal loading of P within the lakes rather than the allochthonous nutrient load. In summer 2012 during conditions of relative drought, we had an opportunity to test whether reduced river P was intensifying P limitation among phytoplankton in impoundments receiving the river water. We compared chlorophyll concentrations and alkaline phosphatase activity measured in 2012 to measurements at the same sites obtained using identical methods in 2004 and 2005, before P reductions were implemented. A large bloom of Aphanizomenon developed in 2012 after a polymictic episode mixed high concentrations of P from the anoxic hypolimnion throughout the lake. Algal biomass was higher and alkaline phosphatase activity levels in 2012 were significantly lower at all sites than in 2004 and 2005, despite reduced supply rate of riverine P. The results confirm that reducing nonpoint source P runoff in the upstream catchment alone is not an efficacious management strategy for these lakes.

Key words: algae, alkaline phosphatase, cyanobacteria, eutrophication, P limitation

Seasonal algal blooms in impoundments of the Huron River sediments rather than those from river inputs was the great- Downloaded by [John T. Lehman] at 05:53 07 June 2013 near Ann Arbor, Michigan, have been recognized as an en- est factor inducing the blooms (Lehman 2011). An exper- vironmental problem for many years. In an effort to reduce imental program of hypolimnetic water discharge prior to nonpoint source loading of phosphorus (P) to the river, the development of anoxia proved to effectively eliminate mid- city of Ann Arbor enacted an ordinance in 2007 banning summer cyanobacteria blooms and replaced them with a the use of lawn fertilizer containing P unless soil test data diatom community (Ferris and Lehman 2007, Lehman et al. identified a specific chemical need for the element (Lehman 2009, Goldenberg and Lehman 2012). The results prompted et al. 2009). Similar legislation was later adopted statewide the municipality with jurisdiction over the lake and its out- in Michigan as Public Act 299 of 2010. Statistical analyses let hydroelectric dam to manage the lake by hypolimnetic after 3 years of post-ordinance observation revealed signifi- withdrawal in subsequent summers, and water quality was cant reductions in river P (Lehman et al. 2011). greatly improved from 2008 to 2011.

Meanwhile, studies in Lake Ford, the Huron River impound- Scientiﬁc interpretations were confounded, however, by the ment historically producing the largest nuisance cyanobac- coincidence that the ordinance restricting lawn fertilizers teria blooms, indicated that internal loading of P from anoxic containing P had been enacted in the upstream drainage in 2007. It remained unresolved whether the P restrictions and reduced allochthonous loading rates had been respon- ∗Corresponding author: [email protected] sible, at least in part, for increased P limitation of lake

116 Alkaline phosphatase

algae and reduced nuisance conditions. Bioassay experi- increased cyanobacteria biomass and reduced P limitation ments performed in 2005 demonstrated that P was the prin- by those algae if the causation theory was correct. cipal element limiting algal biomass development in Ford Lake (Ferris and Lehman 2007), and bioassays in 2008 revealed that Fragilaria achieved dominance over Aphani- zomenon through superior competition for limiting P (Mc- Materials and methods Donald and Lehman 2013). Field site Drought conditions across the Midwest in 2012 prevented Our field site was the middle portion of the Huron River the use of hypolimnetic discharge as a management tool in catchment in southeastern Michigan (United States Geolog- 2012, and Ford Lake again developed cyanobacteria blooms. ical Survey [USGS] Cataloging Unit 04090005; Fig. 1). We took this opportunity to use an existing archive of water We studied the 2 largest impoundments, Ford Lake and chemistry data, including alkaline phosphatase (AP) activity ◦ ◦ Belleville Lake. Ford Lake (42.21 N, 83.56 W) has a sur- rates, to assess the biochemical evidence for P limitation of face area of 4.039 km2 and mean depth of 4.3 m; Belleville post-ordinance algal communities in the absence of whole ◦ ◦ Lake (42.21 N, 83.49 W) has surface area of 3.91 km2 and lake manipulation. Baseline data for AP activity, chlorophyll mean depth of 5.1 m (Nriagu et al. 2001). Both impound- (Chl-a), and water chemistry measured in 2004–2005 (pre- ments were constructed in the early 20th century for hy- ordinance) were compared with data collected by identical droelectric power production, and they continue to produce methods in 2012. We also compared AP activity in 2012 with power commercially. The impoundments are operated as 2006, which was the first year that experimental trials of hy- run-of-the-river, meaning that stage heights are held nearly polimnetic withdrawal were conducted (Ferris and Lehman constant by controlling water flow through turbines and au- 2007). Because AP production is induced by phosphate lim- tomated hydraulic flood gates. itation, the enzyme is a useful quantitative index of P limitation (Healey and Hendzel 1979, Gage and Gorham 1985). Each of the impoundments was sampled at or near its outlet Chrost and Overbeck (1987) found that during summer, AP dam at stations designated F3 and B2 for Ford Lake and activity in the photic zone of a eutrophic lake was mainly Belleville Lake, respectively. Each lake was also sampled produced by phytoplankton rather than bacterioplankton. at an additional site, designated F2 and B1, respectively They also found that AP activity was inversely related to (Fig. 1). The Huron River was sampled at 3 sites where available dissolved inorganic P and positively related to al- runoff was within the jurisdiction of the lawn fertilizer or- gal biomass, measured as adenosine triphosphate (ATP) in dinance, designated A, B, and F, as well as an upstream their study. Other authors have found similar patterns be- site, named CTL, to provide control information about river tween AP activity, inorganic P, and biomass measured as water quality entering the study region (Fig. 1). Chl-a (Labry et al. 2005). With this foundation, our aim was to test 2 apriorihypotheses:

Null Hypothesis: AP activity in 2012 was unchanged from Water sampling and analysis baseline levels measured in 2004 and 2005 despite sig-

Downloaded by [John T. Lehman] at 05:53 07 June 2013 Water samples were collected weekly at each sampling site nificant reductions in P loading into the lakes from the from May to September 2012. Only immediate subsurface river, signifying no change in the P limitation status of (0 m) samples were used for this study. Archival data existed phytoplankton. at weekly or shorter intervals for each site from prior years. Causation Hypothesis: Lake phytoplankton exhibited sig- Again, only immediate subsurface samples were used for nificantly increased AP activity in 2012, signifying in- comparison. Nutrient chemical analyses were identical to creased P limitation within the algal community subse- those described by Ferris and Lehman (2007), Lehman et al. quent to efforts reducing nonpoint source loading of P (2009), and Lehman (2011) and included soluble molybdate- from the upstream catchment. reactive P (SRP), total dissolved P (DP), total P (TP), soluble − molybdate-reactive Si (SRSi), and nitrate (NO3 ). The SRP Application of the scientific method required that we look was measured as molybdate-reactive phosphate in filtrate for the existence of evidence that explicitly should not exist (10 cm path length at 885 nm, 3 replicates, detection limit if the test theories were correct. In this case, the manage- 0.2 μg/L as P); DP and TP were measured as SRP after ment theory being tested was that successful reduction of first oxidizing filtrate (DP) or unfiltered water (TP) with P levels in the Huron River (reduced external loading) had potassium persulfate at 105 C for 1 h. reduced eutrophic conditions and that it both increased the phosphorus deficiency and reduced the biomass of nuisance Samples collected from lake sites were also measured for algae in Ford Lake. We therefore would not expect to find Chl-a and AP activity. For Chl-a, pigment was extracted

117 Lehman et al.

Figure 1.-Site map with river and lake sampling sites identiﬁed.

by macerating Whatman GF/C filters in ice-cold 90% v/v the lake epilimnion through hydroelectric turbines and from acetone by tissue grinder then filtering the slurry through a flood gates at the base of the dam (11 m) were obtained from Whatman GF/D filter. Chl-a was measured fluorometrically logs maintained by the dam operator. using a Turner Designs TD700 fluorometer with 436 nm excitation filter and 680 nm emission filter. AP activity was measured using a modification of Turner Designs application method 998-2679. We added 1 mL aliquots of 36 μM Statistical analyses 4-methylumbelliferyl phosphate (MUP) in 50 mM pH 8.0 All analytes exhibited lognormal distributions and conse- TRIS buffer to 4 mL raw water at 22 C, and both initial and quently were logarithmically transformed prior to statistical time series fluorescence were measured with a Turner De- Downloaded by [John T. Lehman] at 05:53 07 June 2013 analysis. AP activity was the only variable that included val- signs Model 10 fluorometer using the long wavelength UV ues of zero; therefore 1 nM/h (roughly our detection limit) filter kit (P/N 10-302R) with 310–390 nm excitation filter was added to all values before transformation. Univariate and 410–600 nm emission filter. Fluorescence changes were or multivariate analyses of variance (AOV) were performed converted to rate of substrate hydrolysis using reference with IBM SPSS version 19. standards of 2 μM 4-methylumbelliferone as sodium salt. Enzyme activity was expressed as nmol MUP hydrolyzed per hour. Results Precision of measurements Ford Lake sampling and water flow As reported previously (Lehman et al. 2011), the quality YSI temperature sensors and sondes have been moored at control of our water chemistry measurements resulted in F3 since 2008 to provide temperature at 2 m intervals to standard errors <5% of mean values among replicates. AP 8 m depth as well as dissolved oxygen (DO) at 3, 6, and activity, however, proved to be much more intrinsically vari- 8 m every 30 min with data accessible through online data able. Although all assays were performed in triplicate, the radiotelemetry. Prior to 2008, vertical profiles of temperature mean ratio of standard error to sample mean was 0.20 (SD = and DO were measured weekly by boat. Discharge from 0.22, n = 104). Thus, the largest component of uncertainty

118 Alkaline phosphatase

Table 1.-Mean concentrations (and SE) of SRP, DP, and TP (μg/L was similar to 2005 and lower than 2004 and 2006 (Fig. 2), as P) at site F, inlet to Ford Lake from May to Sep of years listed. which are the 3 years when baseline measurements of AP ac- n = number of observations. tivity were made. Mean volumetric retention time from May Years (n) SRP DP TP to September in 2012 (34 d) was almost the same as 2005 (36 d) and considerably longer than 2004 (13 d) and 2006 (17 2004–2005 (44) 19.0 (1.7) 34.4 (2.0) 60.3 (2.8) d). As a consequence of these low flow conditions in 2012, 2008–2010 (66) 16.2 (0.8) 27.8 (1.2) 51.8 (1.4) there was insufficient water available at Ford Lake for the 2012 (21) 16.5 (1.7) 29.9 (2.3) 50.3 (3.9) operator to continuously discharge 300,000 m3/d from the base of the dam, which is the volume required to destabilize the water column and prevent anoxia (Ferris and Lehman in our statistical analyses traces to the AP measurements 2007). Instead, water was discharged mainly from the epil- more so than the associated variables. imnion to generate hydroelectricity (Fig. 3, third panel), and Ford Lake developed and maintained thermal stratification in June and July (Fig. 3, upper panel). Anoxia developed Huron River P below 6 m (Fig. 3, second panel), and when a storm induced Concentrations of P in the Huron River were significantly polymixis in early August, oxygen became uniformly de- reduced from 2008 to 2010 compared with 2003 to 2005 pressed, P concentrations increased (from 2.2 μg/L SRP, as following a ban on the use of lawn fertilizers containing mass of P, on 6 Aug to 43.4 μg/L SRP on 10 Aug), and a P (Lehman et al. 2011). To test whether P concentrations massive bloom of Aphanizomenon flos-aquae developed as in 2012 continued at the reduced level, multivariate AOV stratification reestablished (Fig. 3, lower panel). tests were performed using the analytes SRP, DP, and TP as dependent variables, and period (2008–2010 vs. 2012) and month of year as factors. Our previous research had already established that river P concentrations vary significantly month to month. No significant effect of period AP activity versus specific activity (years) was detected at river sites A and F (P > 0.27), but Although AP is commonly reported as specific activity there were significant differences at sites CTL (P = 0.016) (catalysis rate per unit biomass, e.g., nmol/h/μgChl-a), that and B (P = 0.006). Inspection of the data revealed, however, practice can implicitly suggest a direct, linear relationship that concentrations in 2012 were lower than those from 2008 between enzyme activity and biomass. Such was not the case to 2010 at those 2 sites. Nevertheless, concentrations enter- in this study. The mean slope of ln(AP) versus ln(Chl-a)for ing Ford Lake (site F) were fully comparable to 2008–2010, all sites and years reported in this study was 0.28 (SE = and lower than preordinance conditions (Table 1; P = 0.02 0.14, n = 11; B1 was not measured in 2006), significantly for TP). different from a slope of 1.0 that would indicate linearity in untransformed data. Inside the lakes, however, P concentrations were comparable to the 2004–2005 period. Multivariate AOV of SRP, DP, and Our aprioriexpectation was that AP activity would co- TP detected no significant differences between 2012 and

Downloaded by [John T. Lehman] at 05:53 07 June 2013 vary positively with Chl-a and negatively with SRP; conse- 2004–2005 at F2 (P > 0.9), F3 (P > 0.3), or B2 (P > quently, we evaluated AOV for AP activity at each lake site 0.1). A significant difference was detected only at B1 (P = using SRP and Chl-a as covariates. Our aprioriexpecta- 0.018); inspection of the data revealed that SRP and DP were tion was confirmed in Ford Lake but not in Belleville Lake lower in 2012 than in 2004–2005, but TP concentrations (Table 3). The effect of time period was statistically strong were virtually identical. Concentrations of Chl-a, however, in both lakes, however. Post hoc Bonferroni comparisons of were substantially elevated at all sites in 2012 compared to marginal means revealed that AP activity in 2004–2005 and earlier, preordinance years (Table 2). Chl-a was significantly 2006 were not significantly different but that AP activity was elevated at F3, B1, and B2 (P < 0.03 in each case) but only significantly reduced in 2012 at all sites. A graphic example marginally so at F2 (P = 0.067). Chl-a was also significantly of the data variability and the effects of Period, SRP, and elevated at all sites in 2012 (P < 0.01) compared to 2006, the Chl-a is provided for station F3 (Fig. 4). For comparison first year of experimental selective withdrawal treatments, with other studies, the highest specific rate of AP activity and also a preordinance year. (activity per unit biomass) encountered in this study was 59.1 nmol/h/μgChl-a at station F2 on 11 July 2005; how- River discharge and Ford Lake ever, as presented earlier, specific activity does not represent a suitable variable for statistical tests owing to the nonlinear The year 2012 was marked by drought across the Midwest. effects of independent variables (e.g., both SRP and Chl-a) River discharge measured at Ann Arbor (USGS 04174500) on expressed enzyme activity.

119 Downloaded by [John T. Lehman] at 05:53 07 June 2013 120

Table 2.-Mean concentrations of SRP and TP (as μg/L P) and Chl-a (μg/L) from May to Sep at lake sampling sites for years listed. na = no data available.

2004–2005 2004–2005 2004–2005 Lake–Site SRP 2006 SRP 2012 SRP TP 2006 TP 2012 TP Chl-a 2006 Chl-a 2012 Chl-a

Ford–F2 3.73.93.741.037.7na20.011.835.0 Ford–F3 3.42.73.832.537.036.316.612.328.2 Belleville–B1 5.0na3.651.4na50.622.6na35.4 Belleville–B2 3.62.13.240.140.936.518.116.626.4 Downloaded by [John T. Lehman] at 05:53 07 June 2013

Table 3.-AOV P-values for AP at lake sites, with AP activity as the dependent variable, Period (2004–2005, 2006, or 2012) as an independent factor, and SRP and Chl a as covariates. Marginal means for AP activity, antilog-transformed to nM MUP/h, are listed for 2004–5, 2006, and 2012. NS = not statistically signiﬁcant at α = 0.05; na = not available.

Lake–Site Period SRP Chl-a r2 2004–2005 2006 2012

Ford–F2 0.006 0.001 <0.0005 0.31 136 159 48 Ford–F3 0.003 <0.0005 <0.0005 0.28 136 110 42 Belleville–B1 <0.0005 0.41; NS 0.46; NS 0.25 192 na 61 Belleville–B2 0.001 0.09; NS 0.19; NS 0.23 151 178 46 121 Lehman et al.

Figure 2.-Daily ﬂuvial discharge of the Huron River at Ann Arbor, MI (USGS 041745000) for the 4 years relevant to this report.

Discussion Others have used enzyme-labeled fluorometry (ELF) to in- vestigate variations in AP activity at the cellular level (e.g., Both of our apriorihypotheses were rejected by the ev- Duhamel et al. 2008, 2009, Gonzalez-Gil et al. 1998, Telford idence we collected. Neither was AP activity in 2012 the et al. 2001, Rengefors et al. 2001, 2003, Nedoma et al. 2003), same as 2004–2005 levels, nor did AP activity increase, sometimes using optical microscopy (Dyhrman and Palenik signifying increased P limitation. Instead, AP activity in 1999, Dyhrman et al. 2002) but more commonly using flow both Ford and Belleville lakes was lower in 2012 and algal cytometry. Differences in P limitation status have been iden- biomass measured as Chl-a was higher than that measured in tified both interspecifically and among cells or colonies of any of the baseline years (2004–2005 or 2006) even though a single species; thus, it seems plausible that under compe- incoming river water had lower concentrations of P reflective tition for P, the superior competitors (e.g., diatoms) might of increased control of nonpoint source P runoff. evince less physiological evidence of P stress than would the The results refute both the Null and Causation hypotheses species they outcompeted. The maximum rate of AP activity μ but do not refute the contention that internal loading of P registered per unit Chl-a, 59 nmol/h/ gChl-a, is only 26% Downloaded by [John T. Lehman] at 05:53 07 June 2013 from anoxic sediments is the factor controlling nuisance lower than the maximum specific activity induced by Fer- bloom formation in Ford Lake (Lehman 2011). They also ris and Lehman (2007) in bioassay experiments performed serve as a reminder that efforts to curb P loading from subur- with Ford Lake water incubated for 6 d without added P ban landscapes, while certainly environmentally conscious (Table 5 in Ferris and Lehman 2007) but is considerably and probably salutary in many cases, are not a panacea for less than the maximum Labry et al. (2005) reported for the all lake eutrophication problems. Bay of Biscay (Figure 5b in Labry et al. 2005). Using their assumed stoichiometry of 50 C:1 Chl-a by mass, their maxi- It might seem enigmatic that P limitation as measured by AP mum AP specific activity at zero measurable SRP exceeded activity was not elevated in 2006, the first year in which we 150 nmol/h/μgChl-a. At the opposite extreme, Smith and used whole-lake manipulation by hypolimnetic withdrawal Kalff (1981) reported maximum average June to Septem- to artificially mix Ford Lake to prevent midsummer anoxia; ber AP specific activity from oligotrophic regions of Lake however, our manipulations transformed species composi- Memphremagog (∼10 μg/L TP concentration) of approx- tion of the algal community (Lehman et al. 2009). Bioassay imately 1.5 μg 3-O-methylfluorocein/h/μgChl-a.Given experiments revealed that the diatoms that dominated the the substrate’s formula weight of 346.33, this represents transformed community showed no evidence of P limitation, 4 nmol/h/μgChl-a. In continuous cultures of lake phyto- while growth rates of sympatric subdominant cyanobacte- plankton maintained at low P-limited growth rates, they ria were strongly P limited (McDonald and Lehman 2013). achieved at most 14 nmol/h/μgChl-a (Figure 3 in Smith

122 Alkaline phosphatase Downloaded by [John T. Lehman] at 05:53 07 June 2013

Figure 3.-Vertical distribution of temperature and dissolved oxygen (upper two panels) in Ford Lake from May to Sep 2012. Third panel: daily discharge at Ford Lake dam either from the surface (Qepi) or base (Qhypo). Fourth panel: surface Chl-a at stations F2 and F3. Lower panel: surface TP concentration at F3.

123 Lehman et al.

Figure 4.-Simultaneous ln-transformed concentrations of AP activity, SRP, and Chl-a at Ford Lake station F3 in 2004–2005, 2006, and 2012. Lines are least-square linear regression by year. In the chlorophyll panel, regression lines for 2004–2005 and 2006 are virtually indistinguishable from each other.

Downloaded by [John T. Lehman] at 05:53 07 June 2013 and Kalff 1981). Mean AP-specific activity in Ford Lake anoxia in 2012 (Fig. 3) and high release rates of P from during summer 2012 was 4.7 (SE = 0.8, n = 42) nmol/h/μg anoxic sediments (Lehman 2011) caused Aphanizomenon Chl-a, comparable to Lake Memphremagog, where Smith to be even less P limited that year than during 2004 and and Kalff (1981) did not find phytoplankton to be strongly 2005. P limited. In Lake Champlain, Levine et al. (1997) cited the threshold activity level for severe P deficiency to be 0.2 nmol/min, or 12 nmol/h/μgChl-a. For comparison, mean AP specific activity in Ford Lake during 2004–2005 and Acknowledgments 2006 was 13.0 (SE = 1.2, n = 81) and 11.1 (SE = 1.7, n = 19) nmol/h/μgChl-a, respectively. We thank Mr. M. Saranen and Ypsilanti Charter Township, MI for providing discharge records for the Ford Lake dam as In both 2004–2005 and 2012, the midsummer phytoplank- well as access to real-time data from Ford Lake. We thank ton in Ford Lake was dominated by Aphanizomenon.The J.A. Ferris for assisting with measurements from 2004 to development of extended thermal stratification in 2012 pre- 2006 and K. Karll for assistance with some pigment anal- cluded dominance by diatoms, which have been shown yses in 2012. This research was supported by grants from in this lake to sink quickly from the stratified epilimnion the USEPA, USDA, and the Office of the Provost of the (Ferris and Lehman 2007). It seems that the intense University of Michigan.

124 Alkaline phosphatase

References Lehman EM, McDonald KE, Lehman JT. 2009. Whole lake selective withdrawal experiment to control harmful cyanobacteria Chrost RJ, Overbeck J. 1987. Kinetics of alkaline phosphatase in an urban impoundment. Water Res. 43:1187–1198. activity and phosphorus availability for phytoplankton and Lehman JT. 2011. Nuisance cyanobacteria in an urbanized im- bacterioplankton in Lake Plußsee (North German eutrophic poundment: interacting internal phosphorus loading, nitrogen lake). Microbiol Ecol. 13:229–248. metabolism, and polymixis. Hydrobiologia. 661:277–287. Duhamel S, Gregori G, Van Wambeke F, Mauriac R, Nedoma J. Lehman JT, Bell DW, Doubek JP, McDonald KE. 2011. Reduced 2008. A method for analysing phosphatase activity in aquatic additions to river phosphorus for three years following imple- bacteria at the single cell level using flow cytometry. J Micro- mentation of a lawn fertilizer ordinance. Lake Reserv Manage. biol Meth. 75:269–278. 27:390–397. Duhamel S, Gregori G, Van Wambeke F, Nedoma J. 2009. De- Levine SN, Shambaugh A, Pomeroy SE, Braner M. 1997. Phospho- tection of extracellular phosphatase activity at the single-cell rus, nitrogen, and silica as controls on phytoplankton biomass level by enzyme-labeled fluorescence and flow cytometry: and species composition in Lake Champlain (USA-Canada). The importance of time kinetics in ELFA labeling. Cytom J Great Lakes Res. 23:131–148. Part A. 75A:163–168. Nedoma J, Strojsova A, Vrba J, Komarkova J, Simek K. 2003. Dyhrman S, Palenik B. 1999. Phosphate stress in cultures and Extracellular phosphatase activity of natural phytoplank- field populations of the dinoflagellate Prorocentrum minimum ton studied with ELF97 phosphate: fluorescence quantifi- detected by a single-cell alkaline phosphatase assay. Appl cation and labelling kinetics. Environ Microbiol. 5:462– Environ Microbiol. 65:3205–3212. 472. Dyhrman S, Webb EA, Anderson DM, Moffett JW, Waterbury Nriagu JO, Meier PM, Pascual DL, Lehman JT. 2001. Nutrients JB. 2002. Cell-specific detection of phosphorus stress in in Ford and Belleville Lakes, Michigan, 1998-2000. Report Trichodesmium from the Western North Atlantic. Limnol to Michigan Department of Environmental Quality Surface Oceanogr. 47:1832–1836. Water Quality Division;, August 2001. 99 p. Ferris JA, Lehman JT. 2007. Interannual variation in diatom bloom McDonald KE, Lehman JT. 2013. Dynamics of Aphanizomenon dynamics: roles of hydrology, nutrient limitation, sinking, and and Microcystis (cyanobacteria) during experimental manip- whole lake manipulation. Water Res. 41:2551–2562. ulation of an urban impoundment. Lake Reserv Manage. 29: Gage MA, Gorham E. 1985. Alkaline phosphatase activity and cel- 103–115. lular phosphorus as an index of the phosphorus status of phy- Rengefors K, Pettersson K, Blenckner T, Anderson D. 2001. toplankton in Minnesota lakes. Freshwater Biol. 15:227–233. Species-specific alkaline phosphatase activity in freshwater Goldenberg SZ, Lehman JT. 2012. Diatom response to the whole spring phytoplankton: Application of a novel method. J Plank- lake manipulation of a eutrophic urban impoundment. Hydro- ton Res. 23:435–443. biologia. 691:71–80. Rengefors K, Ruttenberg K, Haupert C, Taylor CD, Howes BL, Gonzalez-Gil S, Keafer BA, Jovine RVM, Aguilera A, Lu SH, Anderson DM. 2003. Experimental investigation of taxon- Anderson DM. 1998. Detection and quantification of alka- specific response of alkaline phosphatase activity in natu- line phosphatase in single cells of phosphorus-starved marine ral freshwater phytoplankton. Limnol Oceanogr. 48:1167– phytoplankton. Mar Ecol Prog Ser. 164:21–35. 1175. Healey FP, Hendzel LL. 1979. Fluorometric measurement of Smith REH, Kalff J. 1981. The effect of phosphorus limitation on alkaline phosphatase activity in algae. Freshwater Biol. algal growth rates: evidence from alkaline phosphatase. Can 9:429–439. J Fish Aquat Sci. 38:1421–1427. Labry C, Delma D, Herbland A. 2005. Phytoplankton and bacterial Telford W, Cox W, Singer V. 2001. Detection of endogenous and Downloaded by [John T. Lehman] at 05:53 07 June 2013 alkaline phosphatase activities in relation to phosphate and antibody-conjugated alkaline phosphatase with ELF-97 phos- DOP availability within the Gironde plume waters (Bay of phate in multicolor flow cytometry applications. Cytometry. Biscay). J Exp Mar Biol Ecol. 318:213–225. 43:117–125.

125